Search Results

The EarLens™ Contact Hearing Device transmits amplified sound by vibrating the eardrum through direct contact. It is indicated for individuals 18 years and older with a mild to severe sensorineural hearing impairment who can benefit from amplification. The device can provide the full spectrum of amplification that includes 125 Hz - 10,000 Hz.

Device Description

The EarLens™ Contact Hearing Device (CHD) transmits amplified sound to compensate for hearing impairment by direct vibration of the tympanic membrane (eardrum). The EarLens™ CHD is composed of an external Audio Processor, which includes a Behind-the-Ear (BTE) Unit and an Ear Tip, a Tympanic Membrane Transducer (TMT), the EarLens™ Fitting Software (ELF), and a Charger with a Power Adapter. In this device, light is used to wirelessly transmit both signal and power from the Audio Processor to the TMT. The BTE sits behind the outer ear, housing the rechargeable battery, digital signal processor (DSP), microphones and drive electronics. The Ear Tip contains the light emitter and directs the light signal down the ear canal. The TMT resides at the end of the ear canal on the skin around the tympanic membrane. The TMT receives the light signal and converts it into direct vibration of the umbo of the tympanic membrane. The EarLens™ Charger charges two BTEs at the same time when connected to either the wall power adapter or from the internal battery contained in the Charger. The CHD is patientmatched for single patient use. The ELF enables the hearing professional to program the device specific to the patient's hearing needs. The EarLens™ Impression System is provided to the physician to enable the collection of a deep ear canal impression to create patient-matched TMTs for each patient.

AI/ML Overview

The EarLens™ Contact Hearing Device (CHD) has undergone extensive testing to ensure its safety and effectiveness. Both non-clinical (bench) studies and clinical studies were conducted to support its performance and establish acceptance criteria.

The acceptance criteria and reported device performance are as follows:

1. Table of Acceptance Criteria and Reported Device Performance

Test Category	Acceptance Criteria	Reported Device Performance
Biocompatibility	No cytotoxicity, skin sensitization, or irritation per ISO 10993 standards. Levels of Nickel leaching within safe European standards (EN1811).	All tests passed. No nickel leaching was noted, and traces of oxidation (discoloration) after prolonged wear were contained beneath the parylene coating, posing no safety risk.
Shelf Life/Sterility	Expected service life of 1 year. Device supplied non-sterile.	Expected service life of 1 year demonstrated through mechanical and reliability testing. Device provided non-sterile.
Electromagnetic Compatibility (EMC)	Max signal variation less than 3dB during EMC tests. Immunity to specified levels for home environment (ESD: +/- 8kV contact, +/- 15kV air; Power frequency magnetic fields: 30 A/m; Conducted RF: 3V r.m.s outside ISM, 6V r.m.s in ISM bands; Radiated immunity: 10 V/m 80 MHz - 2.6 GHz). Performance maintained within +/-3dB in close proximity to intentional radiators.	All EMC tests passed with no variation of more than 3dB observed, demonstrating adequate performance. All immunity tests passed at higher test levels, supporting use in the home environment and on aircraft (RTCA DO-160 Section 20 Category T). No degradation of performance in close proximity to intentional radiators.
Software Level of Concern	Minor, with failures unlikely to cause injury. Maximum Equivalent Pressure Output (MEPO) appropriately limited.	Level of Concern identified as Minor, deemed reasonable given MEPO limitations and hardware component selection.
Electrical Testing	Output Limiter Test: Battery current limited to 12mA ± 1mA at 4.2V.	Passed.
	Harmonic Distortion Test: Distortion less than 5%.	Passed.
	Maximum Output Test (MEPO): Average MEPO ≤ 132 dB SPL.	Average MEPO on four temporal bone samples was 127dB SPL. Passed.
Mechanical & Reliability Testing	Ear Tip Pull Test: Withstands 8oz pull force for 1095 cycles.	Passed.
	Ear Tip Bend Test: Withstands 150° bend for 1095 cycles.	Passed.
	Ear Tip Twist Test: Withstands 150° twist for 1095 cycles.	Passed.
	Accelerated Aging Test (TMT): No performance degradation in TMT output and adhesive bond strengths after 1 year simulated aging at 65°C.	Passed.
	Accelerated Aging Test (Ear Tip): Less than 10% degradation in laser output. No degradation in adhesive bond strength when Ear Tip cable subjected to 10N force for 26 cycles after 1 year simulated aging at 75°C.	Passed.
	Maximum Force on Umbo: Maximum force exerted on the Umbo is ≤ 6mN.	Passed.
	Grasping Tab Mechanical Strength: Withstands pull forces of at least 1N.	Passed.
	Umbo Platform Mechanical Strength: Withstands forces of at least 0.25N.	Passed.
MRI Safety	Device must be labeled "MR Unsafe" and patient provided with a card for TMT removal prior to MRI.	Device is labeled "MR Unsafe" and patient card is required.
Clinical Safety (Primary Endpoint)	No decrease in hearing sensitivity of more than 10 dB in PTA4 (500, 1000, 2000, 4000 Hz) after 4 months of device usage.	No decrease in hearing sensitivity of more than 10 dB was observed. No subject exhibited a >10 dB decrease at any frequency.
Clinical Safety (Secondary Endpoint - AEs)	All AEs temporary and resolved without sequelae. No serious device/procedure-related AEs. No overheating of the ear canal.	31 AEs reported in 20 subjects/22 ears. All but one (sensation of fullness, unresolved but effectiveness not impacted) were temporary and resolved. No serious AEs. Ear canal temperature rise < 1°C.
Clinical Safety (TMT Damping)	Minimal and reversible damping of residual hearing when TMT is in place but BTE unactivated. Average PTA damping ≤ 4 dB.	Average PTA damping was 4 dB, immediately reversible upon TMT removal. Maximum damping of 20 dB at a single frequency. Effect remained constant over 4 months.
Clinical Effectiveness (Primary Endpoint)	Statistically significant improvement in mean aided word recognition (NU-6) at 45 dB HL at 30 days post-placement compared to unaided baseline.	Average baseline unaided score was 52%, average aided score was 85%. Improvement was statistically significant (p<0.0001) and clinically meaningful (33% improvement). Gains persisted through 120 days.
Clinical Effectiveness (Secondary Endpoint - Functional Gain)	More than 10 dB functional gain in thresholds over 2000-10000 Hz for the aided condition at 30 days post-placement compared to unaided baseline.	Mean functional gain was 30.5 dB (p<0.0001), with a maximum of 68 dB at 9-10k Hz. Gains persisted through 120 days.
Clinical Effectiveness (Speech in Noise)	Statistically significant benefit from directional microphone.	Directional microphone provided an average of 3.14 dB benefit over unaided baseline (p<0.0001). Omni-directional mode showed a trend toward improvement but not statistically significant benefit.
Clinical Effectiveness (Perceived Benefit - APHAB)	Statistically significant improvement for EarLens™ relative to unaided.	92% of subjects perceived a statistically significant improvement relative to unaided. Global APHAB scores for EarLens™ were essentially the same as subjects' own hearing aids (1% difference).
Clinical Effectiveness (Patient Satisfaction)	Majority of subjects satisfied with reported outcomes (e.g., sound quality, speech in noise, quality of life).	90% satisfied/very satisfied with speech in noise. 80% improved/very much improved in group conversations. 87.5% improved/very much improved in effort for conversations. 90% satisfied/very satisfied with sound quality. 77.5% improved/very much improved in overall quality of life.

Study Information:

2. Sample Size and Data Provenance:

Pivotal Clinical Study: 48 subjects (96 ears). Prospective, single-arm, open-label study. Subjects served as their own controls (unaided vs. aided). Data provenance is not explicitly stated by country, but the inclusion criteria mention "Native speaker of American English," suggesting the study was conducted in the US.
Ear Impression Study: 78 subjects (154 ears). Retrospective or prospective nature not explicitly stated, but implies observation of a procedure. Again, country of origin is not specified but use of American English would suggest a US-based study.
"Roll-in cohort" (Laser Safety): 10 ears (5 subjects). Prospective.

3. Number of Experts and Qualifications for Ground Truth - Test Set:

For the clinical studies (Pivotal and Ear Impression), the nature of the "ground truth" is based on objective audiometric measurements (hearing thresholds, word recognition scores), subjective questionnaires (APHAB, Study Exit Questionnaire), and clinical observations (tympanic membrane images, adverse event reporting, ear canal temperature measurements).
The raw data for these measurements and observations would be collected by trained healthcare professionals (audiologists, ENT physicians, study coordinators).
Tympanic membrane images (n=78 ears) review: Unspecified number of experts or their qualifications for reviewing these images were provided.
Adverse Event reporting: Medical professionals (investigators, clinicians) determined the nature and resolution of AEs. Their specific qualifications (e.g., radiologist with X years of experience) were not detailed, but they would be medical doctors and/or audiologists involved in the clinical study oversight.
Audiometric data: Audiologists performed the audiometric tests, but the number of experts for establishing "ground truth" (e.g., by consensus on diagnoses) is not specified, as it's typically based on established audiometric thresholds.

4. Adjudication Method for the Test Set:

No specific formal adjudication method (e.g., 2+1, 3+1) is described for the clinical study endpoints or safety events. Results were reported based on statistical analyses of aggregated data and clinical observations by the study investigators.

5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study:

No direct MRMC comparative effectiveness study was mentioned where human readers' performance with and without AI assistance was evaluated. The clinical study compared the device's performance (aided condition) to the unaided condition (subjects essentially serving as their own control). It also made a brief comparison of APHAB scores between the EarLens™ CHD and "subjects' own conventional hearing aids," but this was not a structured MRMC study designed to measure the improvement effect size of human readers with AI vs. without AI assistance. The APHAB comparison showed essentially no clinically significant difference (1%) between the EarLens™ CHD and conventional hearing aids.

6. Standalone (Algorithm Only) Performance:

The device itself is a hearing aid, not an AI-driven diagnostic algorithm. The "performance testing" described in the bench studies (e.g., electrical, mechanical, reliability) assesses the device's physical and functional characteristics in isolation (standalone) from patient interaction or expert interpretation, according to engineering and safety standards.
The "software" component is for fitting and control, not standalone diagnostic or interpretative performance. The clinical study measures the overall system performance (device + patient interaction).

7. Type of Ground Truth Used:

Clinical Studies:
- Safety: Residual hearing function (PTA4 thresholds), adverse event reporting, tympanic membrane images, ear canal temperature measurements.
- Effectiveness: Objective audiometric data (word recognition scores, functional gain/thresholds), subjective patient-reported outcomes (APHAB, Study Exit Questionnaire), speech-in-noise tests (HINT, Quick SIN).
Bench Studies: Established by adherence to national/international standards (e.g., IEC, ISO, ASTM, EN) and internal engineering specifications, using calibrated instruments and measurement techniques.

8. Sample Size for the Training Set:

There's no explicit mention of a "training set" in the context of an Artificial Intelligence (AI) or machine learning model for diagnostic or prognostic purposes, as the EarLens™ CHD is a medical device for hearing amplification rather than an AI diagnostic tool.
The software mentioned (EarLens™ Fitting Software - ELF, and DSP firmware) is for controlling device parameters as programmed by a hearing professional, not for learning from data.
The "design criteria" for the device would have been developed based on general principles of audiology, psychoacoustics, and engineering, likely informed by existing data and research in hearing impairment and hearing aid design.

9. How Ground Truth for Training Set was Established:

Given that this is not an AI/ML-driven diagnostic device with a "training set" in the conventional sense, the concept of "ground truth" for a training set does not directly apply. The device's "training" or development was based on established scientific principles, engineering standards, and the expertise of hearing professionals. The software's "ground truth" is its programmed function to control the device based on clinician input, verified through functional testing against requirements.

Ask a Question

Ask a specific question about this device

Page 1 of 1